Goldman Sachs is forecasting a 7-percentage-point decline in Big Tech return on equity, citing AI capital expenditure that is outpacing any measurable return on investment.1
The bank published a report on AI spending and profitability that doubled as a warning to institutional investors: AI infrastructure buildout is compressing margins before revenue uplift has arrived.1
Return on equity — the ratio of net income to shareholder equity — is the clearest measure of capital efficiency. A 7pp decline signals that AI investment is consuming capital faster than it generates profit. For large-cap technology companies already operating at high valuations, that dynamic creates material downside risk.1
The concern is structural, not cyclical. AI infrastructure requires sustained, multi-year capital commitments: data centers, chips, power infrastructure, and engineering headcount. Revenue from AI products — whether enterprise software, API access, or consumer subscriptions — accumulates more slowly and remains unevenly distributed across the sector.
Goldman's simultaneous publication of a spending report and a ROE forecast is itself a signal. It reflects a shift in how institutional analysts are framing AI: less as a growth catalyst and more as a capital allocation stress test.1
The near-term implication is stock derating. As institutional investors price in margin compression from AI spending cycles, Big Tech AI infrastructure names face valuation pressure even if underlying business fundamentals remain intact.1 Price-to-earnings multiples that were justified by AI growth expectations become harder to defend when ROE is declining.
For corporate finance teams inside these companies, the pressure is mounting to demonstrate returns. Internal hurdle rates, project prioritization, and capital allocation decisions are all being scrutinized against a backdrop of rising debt servicing costs and shareholder expectations that AI investment pays off within a foreseeable window.
The Goldman report does not suggest AI spending will stop — competitive pressure ensures it continues. What it signals is that the market's tolerance for open-ended AI capex without visible returns is narrowing. Investors are beginning to demand accountability metrics: revenue per dollar of AI infrastructure deployed, margin trajectory, and payback period.
That shift in investor posture, from endorsing AI spending to interrogating it, is itself a meaningful change in the capital markets backdrop for technology companies.1
